Yes, we can very quickly build a global Voice-mail and E-mail network, to be used in remote areas from the bottom up,  using available technology.  Low threshold for the users because it works like an answering machine and low investment  to initialise, because it does not need billions of dollars to build an instrastructure for telecom or Internet first.  We think such a program can indeed help to span the digital divide , as the recent G8 meeting in July 2000 pointed out that urgent measures should be taken to establish more "inter- and intraregional networks" in developing countries so they do not fall behind in access to ICT tools and communication networks for culture, commerce and local logistics.

This paper may start a third worldwide avalanche, after the rapid spreading of user access to the Internet and of handsets forWireless Telephony. This time however the avalanche must roll in the horizontal plane and it will maybe even have to roll uphill. Therefore we need your help and loud noises to get it moving.

The proposal is to start definition, design and building of an Email & Voice message network accessible for each and every human on this planet, in every location wherever they are, starting in the most remote places in the world. This open network will use wireless links in parallel in such a way that the wireless  terminal devices are their own repeaters and routers for person-to-person messages. In other words, the communication devices are the infrastructure themselves, which can be grown from two devices upwards and the system can scale up. They will interconnect and interwork with the Internet and the telecom networks and all their telecom networks and mobile devices. So the TELLET Project is an extension and boost to those networks as well. The central ideas behind this tellet proposal were published earlier in 1999 in TelecomMagazine (NL) [1].

1. Introduction

It must be made clear that none of the functions and none of the technical implementations stated in this paper are in themselves new or taken much further than now feasible. Indeed most powersourcing, hardware, software and telecom gear are commercially available and operational right now. New is the functional specification and rough draft system architecture to combine these components in such a way that a new worldwide network may be built, will grow and will be useful.

2. What is the problem?

In all earlier introductions of broadcasting- and telecommunication networks the first users, early adopters were wealthy business persons in urban centres who needed it for their work and could afford to pay for their use as launching customers. Thus commercial ventures could build the necessary telecom infrastructures and extend the reach of the respective local networks. Examples are networks for mail, telegraph and telex , radio broadcasting, telephony, television broadcasting, data communication and fax. In the second (rollout) stage governments realised that further spreading of the connectivity to other parts of the population and to more rural areas could only be implemented in a closed national system of cross-subsidies from rich to poor, from business to population, in the interest of all, since networks would thus become more useful when they could be rolled out to reach more people (Metcalfe Law: network usefulness ~ N^2 if N is the number of subscribers).
Internet and mobile telephony are more recent and have very quickly gone through both stages (introduction and rollout) in most northern countries already. Even the third stage (saturation + diverse better qualities of service) is reached or entered soon in the wealthy G-7 countries and large cities of the world. In develloping countries and especially their rural areas however neither access to fixed / mobile phone or  Internet wired PCs can be seen expected to be avialable in years to come, although a number of organisations (ISOC, UN, IICD) and NGO's are very active to bring communication and knowledge access to villages and schools.  Sometimes by universities and ISPs through VSAT links, or by wireless links on for instance library busses).

The good news is that the growth of the networks everywere on the globe is exponential (fixed doubling time of the number of people with communcation access) and at about the same growth rate (same angle of growth lines, shifted in time, when pictured on log scale). The bad news is however that if we count the number of people who have access to a telephone or to an Internet-enabled PC, in regions or countries of the world, the inequality of the access densities is staggering. Not only between countries but also between urban and rural areas. A popular example is the fact that in the cityarea of Tokio there are more fixed telephone connections ( 20 million ) than in the whole continent of Africa (18 million), and most of these in Africa are either in the large cities north of the Sahara or in South Africa (5 million), see http://www.itu.int/ti/industryoverview/index.htm

A large percentage of the world population at present is not within reach of one day travel of a telephone device or has not yet made a phonecall in their lifetime. The ITU and more recently the UN, have depicted this inequality and have urged governments and business to do more to expedite the deployment of networks in the developing countries which they support. The World Bank and ISOC have pointed to the importance of investing in telecommunication and Internet because of the commercial and social benefits they bring by lower transaction costs and by making knowledge accessible. An added problem is that inhabitants of rural areas now have to travel to district centers or cities to trade or receive care from specialists. This makes life in rural areas or small villages less attractive or even unsustainable and it makes large cities all over the world grow and become more overcrowded every day. Telecom could restore by distance independent-connectedness the feasibility of small villages, including the interrelations and cultural weaving among and between the local inhabitants themselves.

3. Analysis of the problem

If we look more closely at the mechanisms and economics of the deployment of telecom networks we can recognise several processes which can expedite these woven fabrics or are obstacles.

• The huge level of investments which are needed in fixed line and/or wireless telecom infrastructures also needed for Internet access dictate a sequence of priorities of foreign and national investors. Most interests are to obtain concessions to invest and exploit, in order of ROI:  International transmission links, large city switches, national backbones, district city switches and extensions, urban GSM licences and only then.... rural extensions and callboxes. Consequently from arteries/highways to remote farms the interests and financial yields deminish and the total costs to dig trenches and lay cables or hang cables on poles will rise. Moreover, the remote link cables have a higher risk to be stolen. Mobile phone networks are made available in large cities first for those who can afford its cost of usage.

• Satellite systems are expensive to use, are again targetted primarily at rich traveling businessmen and do probably not scale for large numbers of users, even for the new versions which are designed not for telephony but for IP traffic.

• From our experiences with the introduction and rollout of Email (i.e. Fidonet) and Internet it is clear that a specific sequence-pattern of adopting groups is repeated everywere. Scientists use it first and can explain the investments in their networks based on the position they have in the eyes of governments (for instance the nuclear scientists in the former USSR). After them other 'knowledge workers' and professionals in administrations, research and business can get access. Later ISPs for a wider audience start to appear, manned with the young guys and girls who learned to operate internet services at the universities. Etc. Recent studies show the strong correlation between growth in wealth of countries and the way they treat ISPs.

The sequence of subsequent adoption means that effective aid to the process depends on the specific local phase of obstacles, demand and knowledge, just like the (non-linear dynamics) Liebig Law in agriculture does to boost crop yield. Therefore succesfull help to spread usage of Internet is very much depending on local circumstances and levels of knowhow and support.

• National PTT's do not welcome Internet because they jealously guard the earlier mentioned cross-subsidy process to roll out countrywide phone systems.

•  PC's are very expensive to buy and to maintain in the sense of rapid obsolecence and turnover of new software versions, updating of network managenent, user support and costly new components and peripherals. Realistic estimates for LAN- and WAN-network enabled workstations in the area of 10,000 dollars/year per seat. This is way above budgets of remote area people not to mention the power- (availability/quality) and workplace- (dust/humidity) requirements of PCs.

• New HTML features of WWW use in Internet and ever larger attachments (i.e. PPT presentations with short videos) to Email messages demand an ever peak capacity of datacom links even while the average (CIR) data rate is still much lower. The data traffic is bursty like on LANs or the bus of the PC itself. This places high demands on bandwidth of transmission infrastructures and peering points between networks and thus on investments and recurring funding to cope with the bandwidth demand growth each year.

• In a large number of complex fabrics of society, including communication networks, a curious phenomenon of  'succes inequality' can be observed. If for instance we make a list of sizes of cities and SORT them in sequential ranking according to the number of inhabitants a few number of huge ones are the first four or five and then the graph dives down very rapidly and has a very long tail of small towns.

The same applies to insurance companies, airport traffic, frequency of use of words or characters (basis for Morse code) in a specific language, access to records on a diskdrive (basis for hashing and cache memory). In fact in such fabrics the ranking is dynamic. Some large ones fall apart and change places, others merge and change places. The graphs of such ranking were first made by Pareto and Zipf and are popularly known as the 80/20 Law. The graphs show a roughly inverse ranking: the relative size (success) of item N is related to 1/N. As a result no.2 is half the size of no.1, no.3 is 1/3 the size of no.1, etc. In practice we know that the best place to settle for a business is at a junction of roads, and because it becomes busy in that place others will build there too. Success attracts success.

This very unequal succes ranking does indeed apply to webpages [2], [3]. Some of them are hit millions of times a day but most of the WWW pages are rarely looked at, even by search engines. Also in rail, car or data networks the hubs grow faster than the spokes. Benoit Mandelbrot did prove (recently republished [4]) the Zipf ranking distribution based on the notion of fractal structures, see [5]. Simply stated: certain (clusters) patterns are repeated at ever smaller sizes. Yes, this applies to backbone networks connecting networks, connecting networks, connecting networks,etc.! Indeed the orders of magnitude difference in telephone- and in Internet access density between rich and poor countries shows a perfect fit to a Zipf ranking.
Please note that we do NOT say that this inequality is fair, or say that it should be like that because it is a natural phenomenon. On the contrary, in our opinion in some cases (not all, like the size of cities, or frequency of disk accesses) it is extremely unfair. It is even worse.
If the mechanisms of complex structures produce such unequal rankings, any action to make them less skewed may result in a similar gross inequality but maybe at most a change in the sequence and cast of top-actors. Only by recognizing and analysing this process we might be able to do something about it if we want.

The remedy is rather simple. Instead of trying to tilt the 1/N graph, let us try to lift the whole graph so that both "the communication-rich get richer AND the communication-poor get RICHER" !! We must implement a new worldwide network which, in addition to the networks already present, can rank the "connectivity-density" per country district in Mbps(CIR)/Km^2 or [R] according to a graph following (1/N + C). In other words: no tax or cross-subsidy from rich to poor, but install/grow a separate basic grass-roots communication network in remote areas giving a constant C connectivity irrespective of investment economics, having its own virtuous cycles to drive rapid upscaling by design.

4. Requirements and specifications for a new Remote Area email Network

After analysis and exploration of various technical possibilities we propose that in a subsequent design the network terminals & nodes should be built according to the following Requirements and Specs (R/S) :
 

• R/S 1. Let us build a wireless network which can be used by all humans everywere on this planet. E-mail from someone in the Gobi desert asking for medical advice for a kid who fell off his horse. A mother who greets her son by Voicemessage three villages away in Africa. Impossible? Science Fiction? We think we can do make this work. With readily available ICT components. But we must make a smart new combination. Also large human energies and enthousiasm must be unleashed to get this avalanche going.

Bottleneck #1 is that in the developing countries there is not enough money, buying-power and knowledge to install, operate and maintain the necessary countrywide hightech network infrastructures and billing systems for Internet.

• R/S 2.  I therefore choose to implement packetradio. Wireless datacom for IP traffic, but in such a way that the terminals work as routers themselves. The more devices there are and the more routers are added the better the network can forward messages, in series or in parallel. The more devices there are in a specific area the ( • R/S 3.) less radio power they automatically they need to use to reach each other and the better spectrum-(re)use can be applied. So, adding devices by users improves performance for all others. A certain amount of devices, for instance installed on mountain tops, must be maintained there in the interest of all users.

• R/S 4. The devices must be waterproof and have buoyancy so that they can be distributed in rivers or during floods if necessary.

Bottleneck # 2 is that in many places there is no, or no dependable, source of electrical energy.

• R/S 5 The terminals, that I call <<#>>  ( pronounce "radiating crossings") , must be powered by solar energy and, to circumvent bottleneck # 1,  must stay operational after installation without the need for network managers/maintainance • R/S 6 : radio-operatorless.

Bottleneck # 3 is that if we would give each <<#>> device an Email address like a telephone, this would require the addressee of a message to go to one specific device in order to receive his or her Email or Voicemessages. Thus

• R/S 7. is that for each person  personal Email  or Voicemessages can be read and answered from every <<#>> where-ever it is located, like the WWW pop-mail services provided by Twigger.com.

• R/S 8, the <<#>> provide easy to use, and self-explaining Email functionality, which can be typed-in/read by semi-literate users in their own language for person-to-person messaging and community "networking", full part of the reach of the total Internet/ mobiledata E-mail cloud. Internet & GSM/UMTS  are thus extended by many hundreds of millions of people.

• R/S 9. to implement very easy to send/recieve Voicemessages.  This is implemented by audiofile-attachement to E-mail messages;  easy adressing for recording/sending and recieving/listening of person-person voicemessages, for instance by identifying people to be adressed with number and/or  colour sequences; durationlimiting & compression of the audiofiles; making the input and output of the Voicemessages compatible with (to & from) Mobile phones ( SMS + voicemail ) and with Internet E-mail/Web audiofile attachements. This makes the Tellet cloud (pronounce Tell-It !!) interoperable with voicefunctions of mobile phones and PC's !

• R/S 10, the cloud of <<#>> must be able to scale up in numbers without running into limitations of spectrum or other resources. This must allow organic growth and will later involve the need for services of telecom network operators and ISPs. Also, around this network extra services are expected to pop up.

Equally important, it should be noted what the <<#>> devices will NOT do:
 

N1: No infrastructure is needed. The multi-user devices are their own infrastructure.

N2. No operating personel needed.

N3. No PCs.  Later it may be considered to connect PDAs and PC's to the <<#>> if massive parallel networking devices allow this type of traffic.

N4. Non broadcast. • R/S 11 It must be possible however to subscribe or unsusbscribe to newsgroups, discussion groups, news pages.

N5 No spamming, no advertisements • R/S 12 spamfilters must be implemented

N6 No lengthy attachments • R/S 13 Length restrictor

N7 No other function than E-mail and Voicemessages: no phone, no WWW.

• R/S 14. The <<#>> devices should be made and constructed from commercially available components like the Palm PDA and digital radio devices like those from Nokia (Rooftop) or like Cisco (Aironet).

• R/S 15 It must be possible later to  connect to the <<#>>devices equipment to measure parameters of climate and/or pollution.

These specs and requirements may look quite simple but it should be understood that the system as a whole must be resilient in the face of very harsh conditions of nature and use. Furthermore we should be able to start simple with low cost prototypes, crates and glue and start a PROCESS of improvement. The worldwide FidoNet was once started like that by the young Tom Jennings. Maybe we can start building a couple of <<#>>s  with a parasol with glued-on solarcells connected to a Palm and a couple of CB radios?

Like every chartered engineer you of course know that a good design for a complex system is a matter of a few othogonal knots, some self-replicating fractals and a couple of learning circles here and there sprinkled with luck and magic. This needs no further explanation I guess. By the way, by "knots" I do mean couplings of self- and general interest. Several of these knots are embedded in the above requirements.

5. Proposal to start the TELLET PROJECT

TELLET means "Transmission Electronique pour Logistique Locale aux Entierre Terre". Pronounced as "tell it". The domain names www.tellet.nl, www.tellet.net and www.tellet.org have been reserved for this project. I invite you to participate in the following phases. We can together do something that was never done before: interconnect all humans. This may change the quality of life for all of us by a network which was built from the grassroots, bottom-up.

PHASE 1
Interested users and interested suppliers of equipment and/or services are warmly invited to comment on this paper and if possible to recommend improvements, changes, additions. These will be studied and implemented in new versions of this paper, naming the person/company who brought the improvement to my attention, together with additions and improvements by the author himself.
Such comments can be sent to mailto:vantill@stratix.nl   mentioning TELLET in the subject line.

PHASE 2
Design and network architecture phase for prototypes and small scale field trial.

PHASE 3
Companies will be invited to tender for the right to build prototypes and test them in the field in several places in the world.

PHASE 4
From the field trials the final design of <<#>> release 1.0 will be made and the outward specs of such devices will be published to establish open standards of network interconnection and use of this new network. After the launch of this network the TELLET project is stopped and its running and improvements is transferred to others.

Donations from companies will not give them any special rights in the TELLET design and procurement process, but information about the donations may be used by such companies for PR purposes.

Please join this initiative. It is our goal to build global brain connections for the rest of us, thousands of millions of us. In the long run this spaceship Earth will only survive if we can together keep life, work and culture in even the remotest small rural village feasible. It can be done. By remote connectivity of the rural inhabitants lateral amongst themselves and with their more affluent relatives elsewhere, they too can prosper we hope. The alternative is that 5 billion people emigrate to large global cities. And they will try to move to your city.

[1] van Till, Werkelijk Wereldwijd Wireless, TelecomMagazine, page 11, June 1999.

[2] Andrei Broder, et al, Graph structure in the web, paper for the 9th International World Wide Web Conference in Amsterdam, 2000; http://www.almaden.ibm.com/cs/k53/www9.final/

[3] Alan Boyle, Web map shows tangles and gaps, - Network not as interconnected as you might think, study shows-, MSNBC, (Bow Tie Theory),  Did appear in:  http://msnbc.com/news/406956.asp?cp1=1  , but this newspage is no longer available. The "Bow Tie Theory" is also briefly explained in:
http://doc.altavista.com/company_info/press/pr051100.shtml

[4] Benoit B. Mandelbrot, Fractals and Scaling in Finance - Discontinuity, Concentration, Risk -,  Springer, 1997

[5]  van Till, Fractanomics, 1999, http://huizen.dds.nl/~vantill/fractanomics.html

This paper is dedicated to mr, Louis D'Aulnis de Bourouill, one of the most brave men during WWII. He was able by sheer audacity, dashing brilliance and unbelievable luck to keep his mobile radio link operational to pass vital information of the NL resistance to London. Most of his brave collegues where caught while transmitting and executed, but Louis survived, outsmarting the Germans. Every day one step ahead of them. Even today the lives of many depend on telecom links. So let's grow millions of links and wire this planet, from the Gobi desert to the Trinidad fisherman villages.

© Copyright 2000 vantill@stratix.nl

Revision history:
  Version  1.4 draft : June 2 , 2000
  Version  2.1 July 9, 2000, Voice messaging added.
  Version 3.1 July 23, reference to G8 digital divide proposals